PFS a surrogate for overall survival in soft tissue sarcoma trials

View on the News

Validation of surrogate endpoints

In evaluating drug efficacy in oncology clinical trials, overall survival is considered the most reliable and meaningful endpoint because it is objective, precise, and easy to measure. However, it requires a large sample size, prolonged follow-up, and may be confounded by postprogression therapies. Given these drawbacks, surrogate markers are especially useful for rare diseases and diseases with effective subsequent-line therapies.

For a practical definition of surrogate endpoints, researchers use criteria that describe the association between surrogate and clinical endpoints at the individual level and trial level as requirements for validation.

In their systematic review, Zer et al. investigated trial-level surrogacy of progression-free survival and response rate for overall survival in advanced soft tissue sarcoma, concluding that PFS and response rate are appropriate surrogates of overall survival. For trial-level surrogacy analysis, the effective sample size is the number of trials, which should be large enough to reliably estimate the correlation between treatment effects on the surrogate and clinical endpoints. Of note in this review, the sample size was 13 trials.

The influence of postprogression survival on the association between PFS and overall survival is important in this analysis. The median postprogression survival was less than 12 months in most of the studies included, and this figure might improve when more effective drugs become available. Longer postprogression survival may translate to a lower correlation between PFS and overall survival.

Another challenge to the surrogacy of an endpoint may come from new treatments in which the mode of action is substantially different from that used to validate the surrogate. Caution is required in generalizing the validation of surrogate endpoints.

In general, time-to-event outcomes assessed at a single time point (for example, 3-month and 6-month PFS) may be misleading. This is also true of the median value of a time-to-event outcome, such as median PFS. For trial-level surrogacy validation, the hazard ratio is the most appropriate measure for time-to-event outcomes.

Efforts to improve the validation of surrogate endpoints are important in this era of personalized medicine and rapid development of oncology drugs. Surrogate markers usually allow for smaller trials and shorter completion times.

Fengman Zhao, Ph.D., is a biostatistician at Dana-Farber Cancer Institute, Boston. These remarks were part of an editorial accompanying the report by Zer et al. (J Clin Oncol. 2016 March 7. doi: 10.1200/JCO.2015.64.3437).



The use of progression-free survival and response rate as surrogates for overall survival were supported by significant correlations between the endpoints, in randomized trials of advanced soft tissue sarcoma.

However, 3-month progression-free survival (PFS) and 6-month PFS were not significantly correlated with overall survival (OS) and were not recommended as surrogates for OS, according to the researchers.

Significant correlations were observed between overall survival and PFS (correlation coefficient, 0.61) and overall survival and response rate (0.51). Correlations between 3-month PFS and 6-month PFS with overall survival (0.27 and 0.31, respectively) were not significant.

“In soft tissue sarcoma, trial design is particularly challenging, owing to the rarity and heterogeneity of this disease. Time-based endpoints including PFS, 3-month PFS, and 6-month PFS are gaining popularity as primary endpoints in phase III RCTs [randomized controlled trials], despite the fact that current data support their use only to screen for effective drugs in phase II trials,” wrote Dr. Alona Zer of Princess Margaret Cancer Centre and the University of Toronto and colleagues (J Clin Oncol. 2016 March 7. Doi: 10.1200/JCO.2016.66.4581).

“Data show that the assessment of outcomes as a single point in time … only rarely mirrors the hazards of the same endpoint. … As such, the odds ratio for 3-month PFS or 6-month PFS likely do not approximate the hazard ratio for PFS, making it difficult to justify the use of these endpoints in definitive phase III trials.”

PFS and overall survival have shown poor correlation in other cancer types, and evidence suggests that survival post progression may influence the association, with weaker correlations at longer survival post progression. The majority of soft tissue sarcoma reports had survival post progression of less than 12 months, likely explaining the high correlation between PFS and overall survival.

The investigators performed a systematic review of 52 randomized controlled trials, published from 1974 to 2014, that included 9,762 patients who received systemic therapy for advanced/metastatic soft tissue sarcoma.

Comprehensive toxicity assessment was included in just 20 studies (47%) and poorly reported in 6 studies (14%). Few studies included quality of life as a secondary endpoint. The authors noted that in the soft tissue sarcoma setting in which the purpose of systemic treatment may be palliation of symptoms, this is a concern.

Over the 4 decades represented by the systematic review, several trends appeared. Overall, a low proportion of studies included intent-to-treat analyses and clearly defined primary endpoints, but these characteristics improved over time. Endpoint selection has shifted away from response rate in favor of time-based events, including PFS, 3-month PFS, and 6-month PFS. Overall survival was the primary endpoint in just 4% of studies. Studies published in the last 2 decades were more likely to be supported by industry (5% vs. 35%).

Dr. Zer and coauthors reported having no relevant financial disclosures.

Next Article: